Methodological considerations with secondary analyses

The use of secondary data sets has begun to receive increased attention in nursing research. Shrinking grant dollars, continuing pressure to produce replicable generalizable research results, and a growing interest in outcomes management and research have all contributed to an interest in examining how data may be used for multiple purposes. Secondary data analysis may offer the practitioner and researcher ready access to large data sets with multiple variables; yet, there are potentially significant methodological concerns that should be well understood and managed in the initial query design. Considered here will be some of the issues regarding the reliability and validity of large data sets used for secondary analysis, including a discussion on maintaining the integrity of those files as they are moved, merged, and transformed. Finally, methods for managing missing data are reviewed.